Supporting information
Crystallographic Information File (CIF) https://doi.org/10.1107/S0108270109055140/sq3230sup1.cif | |
Rietveld powder data file (CIF format) https://doi.org/10.1107/S0108270109055140/sq3230Isup2.rtv |
K2TaF7 (2.613 g) and KF (0.387 g) were placed in a Pt crucible, gently mixed and finally hermetically sealed in. The system was heated to 1123 K in a furnace for 30 min, followed by slow cooling at 1 K min-1 to 973 K and at 5 K min-1 to 473 K. After the system had cooled to room temperature, a white polycrystalline sample was removed and powdered under dry nitrogen atmosphere.
The X-ray powder diffraction pattern was indexed using the program ITO (Visser, 1969) via the positions of 20 diffraction peaks. The accuracy of the lattice parameters was improved by several cycles of LeBail decomposition of the pattern as implemented in the FULLPROF code (Rodriguez-Carvajal, 1993). The systematic absences suggested extinction symbol P–c. A satisfactory structure solution was obtained in the space group P63mc by using the FOX program (Favre-Nicolin & Černý, 2002).
The solved structure was then expanded to P1 symmetry and the atoms moved from the respective special positions and refined using total energy minimization with the VASP code (Kresse & Furthmüller, 1996; Kresse & Hafner, 1993). The main advantage of such an approach is that the contributions of the atoms to the electron density distribution are not weighted by their scattering power, in contrast with X-ray diffraction. Second, this approach provides a remedy to the chronic difficulty of standard powder refinements, which is the lack of information extractable from a powder pattern (see e.g. Smrčok et al., 2007, 2008).
Geometry analysis using the PLATON program (Spek, 2009) revealed that the symmetry of the optimized structure could be increased to P63mc as the r.m.s. deviations between the atomic coordinates of the P1 structural units, expected to be symmetrically equivalent in P63mc, were negligible. Furthermore, the typical deviation of the optimized fractional coordinates from the expected special positions were only ~0.004 in the P1 model. In the final step, the optimized atomic coordinates were transformed back to the P63mc space group and introduced into a Rietveld refinement; three groups (Ta, K and one for all F atoms) of isotropic displacement parameters plus the scale parameter were refined, keeping the profile and atomic parameters fixed. The refined values of the isotropic displacement parameters, i.e. 0.080 (2)Å2 for Ta, 0.090 (3)Å2 for K and 0.091 (5)Å2 for F, do not show any anomalies that would indicate possible problems with the structural model. The overall fit (Rwp = 0.16) is shown in Fig. 1.
Theoretical calculations were carried out using the VASP package (Kresse & Furthmüller, 1996; Kresse & Hafner, 1993). The calculations were based on the density functional theory with periodic boundary conditions (Jones & Gunnarsson, 1989) using the generalized gradient approximation in the exchange-correlation functional (Perdew et al., 1992). The interactions between ions and electrons were described using the projector augumented wave method (Kresse & Joubert, 1999) with a plane-wave cutoff of 400 eV. The optimization of the structure was performed by the method of conjugated gradient in 4k points (Teter et al., 1989; Bylander et al., 1990).
Data collection: X-POW (Stoe & Cie, DATE?); cell refinement: ITO (Visser, 1969); data reduction: X-POW; program(s) used to solve structure: Fox (Favre-Nicolin & Černý, 2002); program(s) used to refine structure: VASP (Kresse & Furthmüller, 1996; Kresse & Hafner, 1993) and Fullprof (Rodriguez-Carvajal, 1993); molecular graphics: DIAMOND (Brandenburg, 2000); software used to prepare material for publication: PLATON (Spek, 2009).
K3TaF8 | Z = 2 |
Mr = 450.25 | Dx = 4.011 Mg m−3 |
Hexagonal, P63mc | ? radiation, λ = 1.78892 Å |
Hall symbol: P 6c -2c | T = 298 K |
a = 8.2533 (4) Å | white |
c = 6.3196 (4) Å | flat sheet, ? × 0.1 mm |
V = 372.79 (3) Å3 |
Stoe Stadi P diffractometer | Data collection mode: transmission |
Radiation source: sealed X-ray tube | Scan method: step |
Ge(111) monochromator | 2θmin = 7.12°, 2θmax = 69.98°, 2θstep = 0.02° |
Specimen mounting: The sample was mounted between two Mylar foils. |
Least-squares matrix: Full | Profile function: Modified Lorentzian |
Rp = 12.725 | 4 parameters |
Rwp = 15.870 | 0 restraints |
Rexp = 12.490 | 0 constraints |
RBragg = 0.1 | Weighting scheme based on measured s.u.'s |
χ2 = 2.560 | (Δ/σ)max = 0.01 |
3144 data points | Background function: Linear interpolation |
Excluded region(s): none | Preferred orientation correction: None |
K3TaF8 | V = 372.79 (3) Å3 |
Mr = 450.25 | Z = 2 |
Hexagonal, P63mc | ? radiation, λ = 1.78892 Å |
a = 8.2533 (4) Å | T = 298 K |
c = 6.3196 (4) Å | flat sheet, ? × 0.1 mm |
Stoe Stadi P diffractometer | Scan method: step |
Specimen mounting: The sample was mounted between two Mylar foils. | 2θmin = 7.12°, 2θmax = 69.98°, 2θstep = 0.02° |
Data collection mode: transmission |
Rp = 12.725 | χ2 = 2.560 |
Rwp = 15.870 | 3144 data points |
Rexp = 12.490 | 4 parameters |
RBragg = 0.1 | 0 restraints |
Geometry. Note that the refinement method does not provide variances of the optimized parameters. The esd's given below are calculated only from the esd's of the lattice parameters. |
Refinement. The number of refined parameters is that in the last cycle. |
x | y | z | Uiso*/Ueq | ||
Ta1 | 0.33333 | 0.66667 | 0.74070 | 0.080 (2)* | |
K1 | 0.15470 | 0.30940 | 0.24310 | 0.090 (3)* | |
F1 | 0.00000 | 0.00000 | 0.00220 | 0.091 (5)* | |
F2 | 0.43900 | 0.56100 | 0.53440 | 0.091 (5)* | |
F3 | 0.33333 | 0.66667 | 1.05470 | 0.091 (5)* | |
F4 | 0.20030 | 0.79970 | 0.81850 | 0.091 (5)* |
Ta1—F2 | 1.9953 (1) | K1—F4iv | 2.6427 (2) |
Ta1—F3 | 1.9844 (1) | K1—F1 | 2.6848 (2) |
Ta1—F4 | 1.9643 (1) | K1—F1v | 2.7517 (2) |
Ta1—F2i | 1.9953 (1) | K1—F4vi | 2.7614 (2) |
Ta1—F4i | 1.9643 (1) | K1—F3vii | 2.8175 (2) |
Ta1—F2ii | 1.9953 (1) | K1—F2 | 2.8869 (2) |
Ta1—F4ii | 1.9643 (1) | K1—F2ii | 2.8869 (2) |
K1—F4iii | 2.6427 (2) | ||
F2—Ta1—F3 | 130.80 | F2i—Ta1—F4 | 78.29 |
F2—Ta1—F4 | 153.70 | F4—Ta1—F4i | 113.96 |
F2—Ta1—F2i | 81.93 | F2ii—Ta1—F4 | 78.29 |
F2—Ta1—F4i | 78.29 | F4—Ta1—F4ii | 113.96 |
F2—Ta1—F2ii | 81.93 | F2i—Ta1—F4i | 153.70 |
F2—Ta1—F4ii | 78.29 | F2i—Ta1—F2ii | 81.93 |
F3—Ta1—F4 | 75.50 | F2i—Ta1—F4ii | 78.29 |
F2i—Ta1—F3 | 130.80 | F2ii—Ta1—F4i | 78.29 |
F3—Ta1—F4i | 75.50 | F4i—Ta1—F4ii | 113.96 |
F2ii—Ta1—F3 | 130.80 | F2ii—Ta1—F4ii | 153.70 |
F3—Ta1—F4ii | 75.50 |
Symmetry codes: (i) −y+1, x−y+1, z; (ii) −x+y, −x+1, z; (iii) −x, −y+1, z−1/2; (iv) x−y+1, x, z−1/2; (v) x−y, x, z+1/2; (vi) −y+1, x−y+1, z−1; (vii) x, y, z−1. |
Experimental details
Crystal data | |
Chemical formula | K3TaF8 |
Mr | 450.25 |
Crystal system, space group | Hexagonal, P63mc |
Temperature (K) | 298 |
a, c (Å) | 8.2533 (4), 6.3196 (4) |
V (Å3) | 372.79 (3) |
Z | 2 |
Radiation type | ?, λ = 1.78892 Å |
µ (mm−1) | ? |
Specimen shape, size (mm) | Flat sheet, ? × 0.1 |
Data collection | |
Diffractometer | Stoe Stadi P diffractometer |
Specimen mounting | The sample was mounted between two Mylar foils. |
Data collection mode | Transmission |
Scan method | Step |
2θ values (°) | 2θmin = 7.12 2θmax = 69.98 2θstep = 0.02 |
Refinement | |
R factors and goodness of fit | Rp = 12.725, Rwp = 15.870, Rexp = 12.490, RBragg = 0.1, χ2 = 2.560 |
No. of data points | 3144 |
No. of parameters | 4 |
Computer programs: X-POW (Stoe & Cie, DATE?), ITO (Visser, 1969), X-POW, Fox (Favre-Nicolin & Černý, 2002), VASP (Kresse & Furthmüller, 1996; Kresse & Hafner, 1993) and Fullprof (Rodriguez-Carvajal, 1993), DIAMOND (Brandenburg, 2000), PLATON (Spek, 2009).
Ta1—F2 | 1.9953 (1) | K1—F1ii | 2.7517 (2) |
Ta1—F3 | 1.9844 (1) | K1—F4iii | 2.7614 (2) |
Ta1—F4 | 1.9643 (1) | K1—F3iv | 2.8175 (2) |
K1—F4i | 2.6427 (2) | K1—F2 | 2.8869 (2) |
K1—F1 | 2.6848 (2) |
Symmetry codes: (i) x−y+1, x, z−1/2; (ii) x−y, x, z+1/2; (iii) −y+1, x−y+1, z−1; (iv) x, y, z−1. |
In the course of our investigation of the compounds formed in the K–Ta–F system we have synthesized the title compound, (I), crystallizing in the binary KF–K2TaF7 system as a congruently melting product of the two components (Boča et al., 2007; Netriová et al., 2009). The existence of this compound has been suggested previously (Efros & Lantratov, 1963; Kovalev et al., 1973), but up till now no structural data have been given.
Since (I) has been obtained only in the form of a fine powder, its structure was solved from laboratory X-ray powder data (Fig. 1). All six atoms in the asymmetric unit were found to be in the special positions of the P63mc space group: atoms Ta1 and F3 in the 2b position, atoms K1, F2 and F4 in the 6c position, and, finally, atom F1 in the 2a position.
The structure is composed of monocapped [TaF7]2- trigonal prisms (Fig. 2) and octahedra of K+ cations, each with an F- anion at the center. The F-centered K6 octahedra share faces to create infinite chains of composition [FK3]2+ along the c axis (Fig. 3). The discrete [TaF7]2- anions are located between the cationic chains. The individual Ta—F bond distances in the monocapped trigonal prisms (Table 1) are in good agreement with the range of 1.919 (3)–1.976 (2) Å found for the similar polyhedron in the low-temperature phase of K2TaF7 (Torardi & Brixner, 1987). The arrangement of the K atoms in (I) is dictated by the presence of the eighth fluoride anion. While in K2TaF7 two independent K+ cations are nine-coordinated (K—F < 3.0 Å) by the F- anions, all of which are shared with the TaV atoms, in (I) the F1 anions bond exclusively to the unique K1 cation and the remaining seven fluoride anions are shared with atom Ta1. The [FK6] octahedra show only a slight deviation from the ideal geometry as the individual K1—F1 bond distances (Table 1) are close to one another and the deviations from the ideal octahedral K—F—K bond angles are less than 2°. These K—F bond distances agree well with the octahedral K—F distances (2.674 Å) in the structure of carobbiite, KF (Wyckoff, 1963). Different bonding conditions in (I) and K2TaF7 have also led to the different K···K separations, which are reduced from 4.08 Å in K2TaF7 to 3.83 Å in (I). To our knowledge, K3TaF8 belongs to a new structure type.
Our structure determination has ruled out the hypothesis of eightfold fluoride coordination of the Ta atoms as identified in the structure of trisodium octafluorotantalate, Na3TaF8 (Hoard et al., 1956). In the Na3TaF8 structure, solved only from the h0l and hk0 projections calculated using film data and not followed by any refinement, the [TaF8]3- groups are suggested to have a configuration of a square antiprism, with the sodium cations surrounded by six fluoride anions forming an octahedron. However, the uncertainities in the choice of the cell, indicating possible problems with pseudosymmetry (i.e. a monoclinic β angle of ca 121°), a rather low number of structure factors and the absence of any refinement throw some doubts on the accuracy of the structure of that compound.